This research, an extension of NIH grant #RO1-GM52126, will be performed primarily in Brazil at the University of Brasilia in collaboration with Antonio F. Pereira de Araujo. As a broad, long-term objective, it intends to contribute in the search for a potential energy function appropriate for ab initio folding simulations of realistric protein models, which would possibilitate the prediction of protein structures from their amino acid sequences. More specifically, an unspecified hydrophobic potential, which has been studied by the foreign collaborator's group in the context of minimalist lattice models, will be combined with a realistic, off-lattice model, with all protein heavy atoms represented explicitly, which was developed and has been used in the principal investigator's group. The main advantage of the hydrophobic potential over more specific energy functions normally used is that it depends on a realtively small number of parameters, that is, the hydrophobicity of each monomer or atom, depending on the model, and, therefore, has a greater chance of being eventually encoded in the sequence of amino acids. This small number of parameters, however, implies that it is much harder, even in lattice models, to find the appropriate combination of parameters for a given native structure. Since the large flexibility of the off-lattice model is expected to aggravate this problem dramatically, sequence independent but physically motivated structural restrictions, mainly in the form of backbone hydrogen bonds, will have to be appied. The main immediate goal of the project is then to investigate to what extent a potential dominated by unspecific hydrophobic interactions, which can be obtained by the approach previously developed in the context of lattice models, can be combined with physically motivated structural restrictions to fold realistic, off-lattice, protein models.